whole new approach, similar looking results

This version of Iteration uses an 8.8 fixed-point custom back buffer for compositing and compresses to a rgb_565 bitmap on the fly for faster display. I also added a bit of statistical antialiasing/softening to spread out the brightest pixels a bit.

 

After a lot of hard work, I found native-software floating point math is about the same speed as my interpreted-software fixed point math (including the use of look-up tables).

Done